The chemical kinetics of the semi-open hydrous pyrolysis system: Time series analysis of lithostatic pressure and fluid pressure

Abstract

As a problem that has plagued geochemists, the perspective of time has long been recognized as an important factor in organic matter evolution and hydrocarbon generation. This article discusses two time series analysis methods based on econometrics, damping trend and autoregressive integrated moving average (ARIMA), to mine the potential value of data and abstract useful information from pyrolysis experiments. The case studies use marine source rocks (depth of 3119–3230 m, in the LH29–2 well) from the Enping Formation in the Baiyun Sag, the deep-water area (depth of >300 m) of the northern South China Sea, which has a geological background of high temperature, high pressure and rapid burial in the late stage. A tubular plug flow microreactor is used to artificially mature the source rock in the laboratory, and the conditions of hydrous pyrolysis are listed as follows: temperatures of 250°C, 300°C, 350°C, 400°C, 450°C, and 500°C; lithostatic pressure of 33~89 MPa; fluid pressure of 16~38 MPa; and time duration of 72 h. The results revealed that as the degree of maturation increased (Ro from 0.27 to 2.04), the cumulative yield of hydrocarbon gas increased rapidly (max = 311.64 ml/g⋅TOC), the yield of the expelled oil increased first and then decreased (peak = 350.69 mg/g⋅TOC at 350°C), and the yield of residual oil extracted from residue decreased gradually. Although the kinetic parameters of kerogen can be calculated by conventional pyrolysis experiments, the long-term cumulative effect of episodic-hydrocarbon expulsion in real conditions is often neglected. The major contributions of our statistical approach for geological applications are described as follows: (1) the recognition of qualitative changes within hydrocarbon generation and expulsion; (2) the characterization of episodic-hydrocarbon expulsion that describes the evolution of the lithostatic pressure and fluid pressure; (3) the disclosure of new influencing factors abstracted from time series models to describe the frequency distribution of hydrocarbon expulsion; and (4) an evaluation of the dependency of new variables.